EAN: 9789811054655

SPRINGERBRIEFS IN STATISTICS ASYMMETRIC KERNE...

This is the first book to provide an accessible and comprehensive introduction to a newly
developed smoothing technique using asymmetric kernel functions. Further  it discusses the
statistical properties of estimators and test statistics using asymmetric kernels. The topics
addressed include the bias-variance tradeoff  smoothing parameter choices  achieving rate
improvements with bias reduction techniques  and estimation with weakly dependent data. Further
the large- and finite-sample properties of estimators and test statistics smoothed by
asymmetric kernels are compared with those smoothed by symmetric kernels. Lastly  the book
addresses the applications of asymmetric kernel estimation and testing to various forms of
nonnegative economic and financial data. Until recently  the most popularly chosen
nonparametric methods used symmetric kernel functions to estimate probability density functions
of symmetric distributions with unbounded support. Yet many types of economic and financial
data are nonnegative and violate the presumed conditions of conventional methods. Examples
include incomes  wages  short-term interest rates  and insurance claims. Such observations are
often concentrated near the boundary and have long tails with sparse data. Smoothing with
asymmetric kernel functions has increasingly gained attention  because the approach
successfully addresses the issues arising from distributions that have natural boundaries at
the origin and heavy positive skewness. Offering an overview of recently developed kernel
methods  complemented by intuitive explanations and mathematical proofs  this book is highly
recommended to all readers seeking an in-depth and up-to-date guide to nonparametric estimation
methods employing asymmetric kernel smoothing.

Aktuelle Preise zur EAN 9789811054655

Produktinformationen zur 9789811054655